基于粒子群优化模糊聚类的旋转机械故障诊断  

Fault Diagnosis of Rotating Machinery Based on Particle Swarm Optimization and Fuzzy Clustering

作　　者：胡方霞[1,2] 谢志江[1] 

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400030 [2]重庆工商职业学院工程技术系,重庆400052

出　　处：《现代科学仪器》2011年第2期16-19,共4页Modern Scientific Instruments

基　　金：重庆市自然科学基金项目(2008BB3179)

摘　　要：提出基于粒子群优化模糊聚类分析的算法,并将其用于旋转机械的振动故障诊断。该算法以模糊C-均值算法(FCM)的聚类目标函数作为粒子群的适应度函数来衡量各聚类中心的优劣,并依据聚类有效性指标自动确定最优聚类数及聚类中心,有效的结合了FCM极易陷入局部最优的缺点以及粒子群算法全局寻优的优点。实践表明,该方法提高了旋转机械故障诊断的准确率,既可正确判断单一故障,又可有效诊断复合故障,从而证明了该算法的有效性。A method of fuzzy clustering optimized by particle swarm algorithm is put forward and used in vibration fault diagnosis of rotating machinery. The method adopts clustering objective function of fuzzy C-means as fitness func tion of particle swarm to measure the good with the bad of cluster centers. And, the best clustering num and clustering centers are automatically got according to clustering validity function. Application showed that the method increases the correctness rate of fault diagnosis for rotating machinery. It can diagnose not only single fault but also compound faults correctly with high reliability. And so, the effectiveness of the method is proved.

关 键 词：旋转机械 故障诊断 粒子群优化 模糊C一 均值 

分 类 号：TH165.3[机械工程—机械制造及自动化]